<div>The contribution of this research work can be divided into two main tasks: 1) implementing this Electronic Warfare Asset Allocation Problem (EWAAP) with the Genetic Algorithm (GA); 2) Comparing performance of Genetic Algorithm to Particle Swarm Optimization (PSO) algorithm. This research problem implemented Genetic Algorithm in C++ and used QT Data Visualization for displaying three-dimensional space, pheromone, and Terrain. The Genetic algorithm implementation maintained and preserved the coding style, data structure, and visualization from the PSO implementation. Although the Genetic Algorithm has higher fitness values and better global solutions for 3 or more receivers, it increases the running time. The Genetic Algorithm is around (15-30%) more accurate for asset counts from 3 to 6 but requires (26-82%) more computational time. When the allocation problem complexity increases by adding 3D space, pheromones and complex terrains, the accuracy of GA is 3.71% better but the speed of GA is 121% slower than PSO. In summary, the Genetic Algorithm gives a better global solution in some cases but the computational time is higher for the Genetic Algorithm with than Particle Swarm Optimization.</div>

FundaÃÃo Cearense de Apoio ao Desenvolvimento Cientifico e TecnolÃgico / The infrastructure expansion planning in cellular networks, so called Base Station Placement (BSP) problem, is a challenging task that must consider a large set of aspects, and which cannot be expressed as a linear optimization function. The BSP is known to be a NP-hard problem unable to be solved by any deterministic method. Based on some fundamental assumptions of Long Term Evolution - Advanced (LTE-A) networks, this work proceeds to investigate the use of two methods for BSP optimization task: the Particle Swarm Optimization (PSO) and the Differential Evolution (DE), which were adapted for placement of many new network nodes simultaneously. The optimization process follows two multi-objective functions used as fitness criteria for measuring the performance of each node and of the network. The optimization process is performed in three scenarios where one of them presents actual data collected from a real city. For each scenario, the fitness performance of both methods as well as the optimized points found by each technique are presented. / O planejamento de expansÃo de infraestrutura em redes celulares à uma desafio que exige considerar diversos aspectos que nÃo podem ser separados em uma funÃÃo de otimizaÃÃo linear. Tal problema de posicionamento de estaÃÃes base à conhecido por ser do tipo NP-hard, que nÃo pode ser resolvido por qualquer mÃtodo determinÃstico. Assumindo caracterÃsticas bÃsicas da tecnologia Long Term Evolution (LTE)-Advanced (LTE-A), este trabalho procede à investigaÃÃo do uso de dois mÃtodos para otimizaÃÃo de posicionamento de estaÃÃes base: OtimizaÃÃo por Enxame de PartÃculas â Particle Swarm Optimization (PSO) â e EvoluÃÃo Diferencial â Differential Evolution (DE) â adaptados para posicionamento de mÃltiplas estaÃÃes base simultaneamente. O processo de otimizaÃÃo à orientado por dois tipos de funÃÃes custo com multiobjetivos, que medem o desempenho dos novos nÃs individualmente e de toda a rede coletivamente. A otimizaÃÃo à realizada em trÃs cenÃrios, dos quais um deles apresenta dados reais coletados de uma cidade. Para cada cenÃrio, sÃo exibidos o desempenho dos dois algoritmos em termos da melhoria na funÃÃo objetivo e os pontos encontrados no processo de otimizaÃÃo por cada uma das tÃcnicas

Submitted by Rosivalda Pereira (mrs.pereira@ufma.br) on 2017-08-11T19:29:47Z No. of bitstreams: 1 SidneyCerqueiraJunior.pdf: 2395617 bytes, checksum: cbf8e82ed5431d78b69640d4a6b7d511 (MD5) / Made available in DSpace on 2017-08-11T19:29:47Z (GMT). No. of bitstreams: 1 SidneyCerqueiraJunior.pdf: 2395617 bytes, checksum: cbf8e82ed5431d78b69640d4a6b7d511 (MD5) Previous issue date: 2012-06-01 / In the last years, the process of restructuring of the electricity market, brought several changes in the operational e regulatory aspects. The main idea was the separation of the generation, transmission and distribution activities in order to insert the competition among them, aimed to increase the e ciency, safety and quality of supply of electrical energy. The hourly schedule, usually called a Unit Commitment has as objective the de - nition of which generators should be online/o ine and their respective operation points. In some markets based on this new model, the determination of the optimal scheduling of generators (thermal and hydro) is made by the Agent Generator, which is largely responsible for the allocation of your portfolio. Given this, the aim of this work is to nd the operational policy that will maximize the pro t of Agent Generator, based on forecast price and respecting the thermal, hydro and market constrictions assigned to the problem. Thus, the optimal schedule found is an important factor in developing strategies to o ers of bids to auctions in which the Genco will participate. For the case study technique Particle Swarm Optimization is applied to solve the problem in plants belonging to the Brazilian electric system, which are also analyzed the in uence of the start-up cost to the optimal schedule. / Nos últimos anos, o processo de reestruturação da indústria da eletricidade, trouxe diversas mudanças nos aspectos operacionais e regulatórios. A ideia principal foi a separação das atividades de geração, transmissão e distribuição, de modo a inserir competição entre esses, visando o aumento da e ficiência, segurança e qualidade no fornecimento da energia elétrica. A programação horária, usualmente denominada de Pré-Despacho de Potência, tem como objetivo a defi nição de quais unidades devem estar ligadas/desligadas e seus respectivos pontos de operação. Em alguns mercados baseado neste novo modelo, a determinação da programação ótima dos geradores (termelétricas e hidrelétricas) é feita pelo próprio Agente Gerador, sendo este o maior responsável pela alocação de seu portfólio. Diante disto, o objetivo deste trabalho é encontrar a política operativa que irá maximizar o lucro desse Agente Gerador, baseado na previsão de preço horário e respeitando as restrições térmicas, hidráulicas e de mercado atribuídas ao problema. Assim, a programação ótima encontrada é um importante fator para elaboração das estratégias de ofertas de lances a leilões em que o Agente Gerador irá participar. Para estudo de caso, a técnica Otimização por Enxame de Partículas é aplicada para solucionar o problema em usinas que pertencem ao sistema elétrico brasileiro, onde é analisado também a influência do custo de partida na programação ótima horária.

The subject of this thesis is the development of localization algorithms for target localization in wireless sensor networks using received signal strength (RSS) measurements or Quantized RSS (QRSS) measurements. In chapter 3 of the thesis, target localization using RSS measurements is investigated. Many existing works on RSS localization assumes that the shadowing components are uncorrelated. However, here, shadowing is assumed to be spatially correlated. It can be shown that localization accuracy can be improved with the consideration of correlation between pairs of RSS measurements. By linearizing the corresponding Maximum Likelihood (ML) objective function, a weighted least squares (WLS) algorithm is formulated to obtain the target location. An iterative technique based on Newtons method is utilized to give a solution. Numerical simulations show that the proposed algorithms achieves better performance than existing algorithms with reasonable complexity. In chapter 4, target localization with an unknown path loss model parameter is investigated. Most published work estimates location and these parameters jointly using iterative methods with a good initialization of path loss exponent (PLE). To avoid finding an initialization, a global optimization algorithm, particle swarm optimization (PSO) is employed to optimize the ML objective function. By combining PSO with a consensus algorithm, the centralized estimation problem is extended to a distributed version so that can be implemented in distributed WSN. Although suboptimal, the distributed approach is very suitable for implementation in real sensor networks, as it is scalable, robust against changing of network topology and requires only local communication. Numerical simulations show that the accuracy of centralized PSO can attain the Cramer Rao Lower Bound (CRLB). Also, as expected, there is some degradation in performance of the distributed PSO with respect to the centralized PSO. In chapter 5, a distributed gradient algorithm for RSS based target localization using only quantized data is proposed. The ML of the Quantized RSS is derived and PSO is used to provide an initial estimate for the gradient algorithm. A practical quantization threshold designer is presented for RSS data. To derive a distributed algorithm using only the quantized signal, the local estimate at each node is also quantized. The RSS measurements and the local estimate at each sensor node are quantized in different ways. By using a quantization elimination scheme, a quantized distributed gradient method is proposed. In the distributed algorithm, the quantization noise in the local estimate is gradually eliminated with each iteration. Simulations show that the performance of the centralized algorithm can reach the CRLB. The proposed distributed algorithm using a small number of bits can achieve the performance of the distributed gradient algorithm using unquantized data.

The global energy demand in residential sector is increasing steadily every year due to advancement in technologies. The present electricity grid is designed to support peak demand rather than Peak to Average (PAR) demand. Utilities are investigating the residential Demand Response (DR) to lower the (PAR) ratio and eliminate the need of building new power infrastructure. This requires Home Energy Management System (HEMS) at grid edge to manage and control the energy demand. In this thesis, we presented an MDPSO based DR enabled HEMS model for optimal allocation of energy resources in a smart dwelling. The algorithm is designed to lower peak energy demand as well as encourage the active participation of customers by offering a reward to comply with DR request. We categorized appliances as elastic non-deferrable loads and inelastic deferrable loads based on their DR potential and operating characteristics. The scheduling of elastic and inelastic class of appliances is performed separately using canonical and binary version of PSO given how we expressed out load categories. We performed use case simulation to validate the performance of MDPSO for combination of different tariffs: Time of Use (TOU), TOU and Critical peak rebate signal (CPR), TOU and upper demand limit. Simulation results show that algorithm can reduce the electricity cost in range of 28% to 7% under increasing comfort conditions in response to TOU prices and Peak demand reduction of about 24% under TOU pricing and medium comfort conditions for single household. Under CPR DR requests, with respect to TOU pricing, there is effectively no change in the peak under the minimum comfort scenario. Furthermore, algorithm is able to suppress the peak upto 25% under combination of TOU and hard constraint on maximum power withdrawn from grid with no change in the electricity cost. Scheduling of multiple houses under TOU pricing results in peak reduction of 7 % as compared to baseline state. Under combination of TOU and CPR the aggregate peak energy demand of multiple households during DR activation time intervals is reduced by 32 %. The algorithm can suppress the peak demand by 27% under TOU and hard constraint on maximum power withdrawn from grid by multiple houses.

Accurate and fast machine learning (ML) algorithms are highly vital in artificial intelligence (AI) applications. In complex dataset problems, traditional ML methods such as radial basis function neural network (RBFN), sparse coding (SC) using dictionary learning, and particle swarm optimization (PSO) provide trivial results, large structure, slow training, and/or slow testing. This dissertation introduces four novel ML techniques: the multi-column RBFN network (MCRN), the projected dictionary learning algorithm (PDL) and the multi-column adaptive and non-adaptive particle swarm optimization techniques (MC-APSO and MC-PSO). These novel techniques provide efficient alternatives for traditional ML techniques. Compared to traditional ML techniques, the novel ML techniques demonstrate more accurate results, faster training and testing timing, and parallelized structured solutions. MCRN deploys small RBFNs in a parallel structure to speed up both training and testing. Each RBFN is trained with a subset of the dataset and the overall structure provides results that are more accurate. PDL introduces a conceptual dictionary learning method in updating the dictionary atoms with the reconstructed input blocks. This method improves the sparsity of extracted features and hence, the image denoising results. MC-PSO and MC-APSO provide fast and more accurate alternatives to the PSO and APSO slow evolutionary techniques. MC-PSO and MC-APSO use multi-column parallelized RBFN structure to improve results and speed with a wide range of classification dataset problems. The novel techniques are trained and tested using benchmark dataset problems and the results are compared with the state-of-the-art counterpart techniques to evaluate their performance. Novel techniques’ results show superiority over techniques in accuracy and speed in most of the experimental results, which make them good alternatives in solving difficult ML problems.

No description available.

In recent year, the awareness of environmental protection has made the power dispatch problem not necessarily economy-oriented. This thesis proposed the application of Particle Swarm Optimization (PSO) algorithm to solve the Unit Commitment (UC) problem for 24 hours with maximum profit in the power and carbon market. Optimal Power Flow (OPF) is used to solve the UC problem for the interconnected power network that is comprised of three independent areas to optimize the dispatching strategy. The UC problem must satisfy the constraints of the load demand, generating limits, minimum up/down time, ramp rate limits, and also the limits of power flow, buses voltage and transmission line capacity. The other objective of this thesis is to employ the Static Synchronous Series Compensator (SSSC) to integrate with OPF based on Equivalent Current Injection (ECI) power flow model, and install it at interconnected lines between each independent area controlling the power flow to reduce emission. In order to avoid the local optimality problem, this thesis proposed the utilization of the Multiple Particle Swarm Optimization (MPSO), which can quickly reach the optimal solution with a better performance and accuracy. The Independent Power Producer (IPP) can get the maximum profit with installed SSSC from the power and carbon trading with the calculation of power wheeling expense and carbon forecasting data. Furthermore, it can also assess the need of participating in the trading market or not.

In recent years, with the energy shortage, the use of renewable energy is inevitable. With CO2 the most important greenhouse gas causing global warming as well as the increase of population, renewable energy is one way to save energy and reduce carbon emissions. The traditional capacity investment for serving the load in distribution systems usually considered the addition of new substations or expansion of the existing substation and associated new feeder requirement. Nowadays, there are a lots of distributed generations (DG¡¦s) to be chosen. Factors of the choice taken into account will include lower pollution, higher efficiency, higher return rate for construction of distributed power generation systems. This thesis assumes that the distributed generation can be invested for long-term power plant planning. The planning of DG would be investigated from the perspectives of the independent investors. The modified Particle Swarm Optimization is proposed to determine the optimal sizing and sit of DG¡¦s addition in distribution systems with the constrains of CO2 limitation and addition of distributed generation to maximize profits. This thesis deals with discrete programming problem of optimal power flow, which includes continuous and discrete types of variables. The continuous variables are the generating unit real power output and the bus voltage magnitudes, the discrete variables are the shunt capacitor banks and sit problems. The Miaoli-Houlong system of Taiwan power will be used in this thesis for the verification of the feasibility of the proposed method.

Carbon dioxide (CO2) is the most important component of Greenhouse Gas (GHG) that causes global warming and sea-level rising. Thermal power plants dominate electric power generation in the world, and has been reported to be the major contributor of CO2 emission. To prevent the related global warming caused by GHG emission, carbon quota trading is implemented and becomes a gradually arising market. This thesis proposed a research focused on the relationship between the carbon trading scheme and dynamic economic dispatch (DED) problem for the public utility. A model of the carbon trading market was investigated and introduced into DED problem incorporating wind and solar power plant. A refined particle swarm optimization (PSO) algorithm, PSO with time-varying acceleration coefficients (PSO-TVAC), is applied to determine the DED strategy with the incorporation of independent power providers (IPPs) and green power plant. The model of the carbon trading was considered in the DED problem. Carbon reduction is treated as the inner-cost of utility, and the fictitious carbon quotas can be resold to the market, while the energy shortage can be satisfied by purchasing quotas from the market. In order to avoid premature convergence of the original PSO, the PSO-TVAC method is introduced to improve the searching efficiency.